Differential Effects of Pharmacological HIF Preconditioning of Donors Versus Recipients in Rat Cardiac Allografts


  • M. A. I. Keränen,

    1. Transplantation Laboratory, Haartman Institute, University of Helsinki and HUSLAB, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • R. Tuuminen,

    1. Transplantation Laboratory, Haartman Institute, University of Helsinki and HUSLAB, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • S. Syrjälä,

    1. Transplantation Laboratory, Haartman Institute, University of Helsinki and HUSLAB, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • R. Krebs,

    1. Transplantation Laboratory, Haartman Institute, University of Helsinki and HUSLAB, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • G. Walkinshaw,

    1. FibroGen, Inc., San Francisco, CA
    Search for more papers by this author
  • L. A. Flippin,

    1. FibroGen, Inc., San Francisco, CA
    Search for more papers by this author
  • M. Arend,

    1. FibroGen, Inc., San Francisco, CA
    Search for more papers by this author
  • P. K. Koskinen,

    1. Transplantation Laboratory, Haartman Institute, University of Helsinki and HUSLAB, Helsinki University Central Hospital, Helsinki, Finland
    2. Department of Medicine, Division of Nephrology, University of Helsinki and Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • A. I. Nykänen,

    1. Transplantation Laboratory, Haartman Institute, University of Helsinki and HUSLAB, Helsinki University Central Hospital, Helsinki, Finland
    2. Department of Cardiothoracic Surgery, University of Helsinki and Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author
  • K. B. Lemström

    Corresponding author
    1. Department of Cardiothoracic Surgery, University of Helsinki and Helsinki University Central Hospital, Helsinki, Finland
    • Transplantation Laboratory, Haartman Institute, University of Helsinki and HUSLAB, Helsinki University Central Hospital, Helsinki, Finland
    Search for more papers by this author

Corresponding author: Karl B. Lemström



Ischemia-reperfusion injury (IRI) induces hypoxia-inducible factor-1 (HIF-1) in the myocardium, but the consequences remain elusive. We investigated HIF-1 activation during cold and warm ischemia and IRI in rat hearts and cardiac syngrafts. We also tested the effect of HIF-α stabilizing prolyl hydroxylase inhibitor (FG-4497) on IRI or allograft survival. Ex vivo ischemia of the heart increased HIF-1α expression in a time- and temperature-dependent fashion. Immunohistochemistry localized HIF-1α to all cardiac cell types. After reperfusion, HIF-1α immunoreactivity persisted in smooth muscle cells and cardiomyocytes in the areas with IRI. This was accompanied with a transient induction of protective HIF-1 downstream genes. Donor FG-4497 pretreatment for 4 h enhanced IRI in cardiac allografts as evidenced by an increase in cardiac troponin T release, cardiomyocyte apoptosis, and activation of innate immunity. Recipient FG-4497 pretreatment for 4 h decreased infiltration of ED1+ macrophages, and mildly improved the long-term allograft survival. In syngrafts donor FG-4497 pretreatment increased activation of innate immunity, but did not induce myocardial damage. We conclude that the HIF-1 pathway is activated in heart transplants. We suggest that pharmacological HIF-α preconditioning of cardiac allografts donors would not lead to clinical benefit, while in recipients it may result in antiinflammatory effects and prolonged allograft survival.




angiopoietin-related protein 4


carbonic anhydrase IX


cardiac allograft vasculopathy


cytokine induced neutrophil chemoattractant 1


cardiomyocyte; CoCl2,cobalt chloride


cardiac troponine T


Dark Agouti






HIF-α stabilizing prolyl-hydroxylase inhibitor


hyaluron synthase 1


hypoxia-inducible factor-1


heat-shock protein 27


heme oxygenase-1


inducible nitric oxide synthase


ischemic preconditioning


monocyte chemotactic protein-1


nuclear factor-κB


postcapillary venule




prolyl-hydroxylase inhibitors


polymorphonuclear cell


remote ischemic preconditioning


transverse relaxation time


toll-like receptor 4


transforming growth factor-β


tumor necrosis factor-α


terminal deoxyribonucleotidyl transferase-mediated dUTP nick end-labeling


Wistar Furth


Constant shortage of donor organs hinders transplantation activity. Thus, new modalities to increase the available pool of transplants are warranted. Ischemic preconditioning (IPC) refers to brief intermittent periods of ischemia prior to subsequent longer ischemia. Such an IPC protocol is cardioprotective, as it prevents ischemia-reperfusion injury (IRI) [1]. The concept of remote ischemic preconditioning (rIPC) emerged as a brief ischemic insult of the circumflex artery of the heart also protects the adjacent myocardium [2]. rIPC of the kidney or skeletal muscle also reduces myocardial infarct size [3, 4]. The mechanisms of IPC and rIPC remain elusive, although they may involve humoral and neurogenic/neuroendocrine mechanisms [5]. Hypoxia-inducible factor-1 (HIF-1) and HIF-2, key regulators of the physiologic and pathophysiologic response to hypoxia, may be fundamental in the IPC and rIPC processes [6].

In experimental heart transplantation, IPC of the donor heart prevents myocardial stunning [7], and rIPC of the recipient's lower limb ameliorates IRI of the cardiac transplant [8]. Such IPC protocols emphasize the brief and/or intermittent nature of the initial stimuli and short and transient ‘window of opportunity’ for the subsequent ischemia. The achieved protection disappears in the rat heart after just 1 h of subsequent ischemia, and brain death completely abolishes the beneficial effect of IPC [9, 10]. Emerging interest is aimed to generate a prolonged state of cardioprotection via chronic pharmacological preconditioning [11].

Regulation of HIF-1 stability is under tight control of oxygen-sensitive prolyl-hydroxylases (PH) [12]. Inhibition of PH by pharmacological prolyl-hydroxylase inhibitors (PHI) results in HIF-1 stabilization and activation at normal oxygen tension. None of the existing heart transplantation studies use hypoxia-mimicking pharmacological agents, or elucidate the role of HIF-1 in IPC or rICP. Here, we investigated activation of the HIF-1 pathway during graft preservation (cold and warm ischemia) and IRI in native rat hearts and in syngeneic and allogeneic heart transplants. To delineate the therapeutic potential of pharmacological preconditioning, we treated either the allograft donors or the recipients with a novel HIF-α stabilizing PHI, FG-4497 [13].

Our results indicate that HIF activation in allograft donors 4 h prior to heart removal enhances IRI, cardiomyocyte apoptosis and innate immune activation, whereas HIF activation in allograft recipients 4 h prior to heart transplantation reduces inflammation, and mildly improves the long-term allograft survival. The results further suggest that pharmacological HIF preconditioning of cardiac allograft donors would not provide clinical benefit, whereas pharmacological HIF preconditioning of cardiac allograft recipients may have antiinflammatory effects.


Experimental design

Allogeneic cardiac transplantations were performed between fully MHC-mismatched male Dark Agouti (DA, RT1av1 donor and Wistar-Furth (WF, RT1u) recipient rats (Scanbur, Sollentuna, Sweden) weighing 300–350 g [14]. Syngeneic heart transplantations were performed between DA rats. The lighting conditions in our animal unit are standardized to a 6.00 am–6.00 pm room light cycle. All animal experiments were performed between 8.00 am and 4.00 pm. The possible effect of circadian HIF-1 expression on our results should be minimal [15]. Permission for animal experimentation was obtained from the State Provincial Office of Southern Finland. The animals received care in compliance with the ‘Guide for the Care and Use of Laboratory Animals’ (National Academy of Sciences, 2011; ISBN 978-0-3-0-15400-0).

To investigate the effect of cold ischemia on the activation of the HIF-1 pathway during graft preservation, nontransplanted DA hearts were removed for immediate analysis or preserved in heparinized PBS at +4°C for 4 h (n = 6 rats/group). The effect of subsequent warm ischemia occurring during graft implantation was studied by preserving nontransplanted DA hearts in rat peritoneal cavity for additional 45 min (n = 3 rats/group). To investigate the role of IRI per se without the effect of cold ischemic preservation or alloimmune response on the activation of the HIF-1 pathway after heart transplantation, DA hearts were harvested and immediately transplanted to DA rats (total ischemia 1 h). Syngrafts were removed at 15 and 30 min and 1, 4 and 24 h after transplantation and reperfusion (n = 3 rats/group). Quantitative real-time RT-PCR, ELISA and immunohistochemistry measured endogenous HIF-1α.

FG-4497 (FibroGen Inc., San Francisco, CA, USA) is a prolyl-hydroxylase inhibitor which stabilizes HIF-1α and HIF-2α isoforms [13]. FG-4497 was diluted with 1% high-viscosity carboxymethylcellulose sodium and 0.2% Polysorbate 80 (Tween 80). Normal DA rats were pretreated with FG-4497 (n = 4) or left untreated (n = 4). After 4 h of pretreatment the hearts were harvested and subjected ex vivo to 4-h cold preservation, and mRNA levels of HIF-1 downstream genes were analyzed. In the donor treatment group (n = 7), DA rats were treated with a single dose of 60 mg/kg of FG-4497 via nasogastric tube 4 h before heart removal. In the recipient pretreatment group (n = 9), WF rats were treated with a single dose of 60 mg/kg of FG-4497 via nasogastric tube 4 h before heart transplantation. In the control group (n = 10), neither the donor nor recipient rats did receive any additional treatment. The allografts were subjected to 4-h cold ischemia before transplantation and were removed for analysis 6 h after reperfusion. Similar methodology was used to investigate the effect of FG-4497 donor pretreatment in syngenic setting (n = 5). To delineate the effect of donor or recipient pretreatment more thoroughly, we performed allograft survival experiment. The controls (n = 11), the donors (n = 6) and the recipients (n = 10) were pretreated as previously described, and all recipients were treated subcutaneously (SC) with cyclosporine A (CsA; Novartis, Basel, Switzerland) diluted in Intralipid (Fresenius Kabi, Bad Homburg, Germany) at 1 mg/kg/day until the end of the experiment. The pulse of the allograft was measured by daily palpations. The rats were followed until the allograft stopped beating or at 8 weeks. Supporting Materials and Methods give detailed information.


All data were expressed as mean ± SEM and analyzed using SPSS for Mac version 20.0 (SPSS inc., Chicago, IL, USA). We used the nonparametric Kruskall–Wallis with Dunn's test in multiple comparisons and the nonparametric Mann–Whitney U-test in pairwise comparison. For survival, Kaplan–Meier analysis with log rank (Mantel-Cox) was applied. p-values were as indicated and p < 0.05 was regarded as statistically significant.


Preservation injury activates HIF-1α in a time- and temperature-dependent fashion in the rat heart

To investigate HIF-1α activation during cold preservation, Dark Agouti (DA) rat hearts were removed for immediate analysis or preserved in +4°C PBS for 4 h. To mimic the warm ischemia occurring during the heart transplantation procedure, two similar groups of rat hearts were subjected to additional warm ischemia at room temperature for 1 h (Figure 1A).

Figure 1.

HIF-1α is stabilized in time- and temperature-dependent fashion during transplant ischemia. Nontransplanted normal DA hearts were subjected ex vivo to 0- or 4-h cold ischemia with or without 45-min warm ischemia according to the scheme (n = 3–6/group) (A). HIF-1α Western blot analysis from in vitro CoCl2-treated COS7 cells (B). Analysis of nuclear HIF-1α protein stability, as determined by ELISA from nuclear protein extracts (C). Analysis of HIF-1α immunoreactivity in PCV (D), arterial EC (E), SMC (F) and cardiomyocytes (G) during cold and warm ischemia. Representative photomicrographs from the heart cross section show no HIF-1α immunoreactivity in PCV (H), SMC (I), arterial EC (I) or cardiomyocytes (J) in normal hearts. Hearts subjected to 4-h cold ischemia with 45-min additional warm ischemia show HIF-1α immunoreactivity in PCV (K, arrows), SMC (L, arrows), arterial EC (L, arrowheads), and cardiomyocytes (M, arrows). Negative control (isotype-matched IgG) in the upper right corner (M). Representative photomicrographs from the kidney cross section show HIF-1α immunoreactivity in tubular cells after in vivo 30-min warm ischemia (N, arrows) and no immunoreactivity in sham operated (O, arrows). Hematoxylin-eosin background staining. *p > 0.01 **p > 0.05. Scale bars are 20 μm.

Western blot analysis confirmed the specificity of a monoclonal mouse antihuman antibody against HIF-1α, as it detected the expected 120 kDa band in nuclear protein extracted from Cos-7 cells treated with CoCl2 (Figure 1B). We have previously shown that this antibody also detects HIF-1α protein produced after adeno-associated virus-mediated gene transfer of HIF-1α into rat cardiomyocytes in vivo [16].

Analysis of nuclear protein extracts from DA rat hearts by ELISA revealed that prolonged cold preservation increased the levels of nuclear HIF-1α protein in a time-dependent fashion and warm preservation potentiated the nuclear HIF-1α protein stabilization (Figure 1C, p < 0.01).

In immunohistochemical analysis, HIF-1α immunoreactivity was localized time- and temperature dependently in the nuclei of EC of postcapillary venules (Figure 1D, H and K, arrows, p < 0.05), and arteries (Figure 1E, I and L (inset), arrowheads, p = 0.0517), in the nuclei of SMC in coronary arteries (Figure 1F, I and L, arrows, p < 0.01) and in the nuclei of cardiomyocytes (Figure 1G, J and M, arrows, p < 0.05). Representative photomicrographs show low HIF-1α immunoreactivity in normal hearts (Figure 1H–J), while prominent HIF-1α immunoreactivity was achieved after 4 h of cold ischemia with additional 1 h of warm ischemia (Figure 1K–M). Negative controls with isotype matched IgG did not show any immunoreactivity (Figure 1M, inset).

The HIF-1α antibody also detected HIF-1α protein in tubular cells of kidney with 30 min warm in vivo ischemia (Figure 1N), whereas not in the sham operated (Figure 1O). The expression pattern of HIF-1α reflected that of previously described in the literature [17].

HIF-1α immunoreactivity persists in the myocardium during reperfusion injury

To investigate the role of IRI per se without the effect of cold ischemic preservation or alloimmune response on the activation of the HIF-1 pathway after heart transplantation, DA hearts were harvested and immediately transplanted to DA rats. Syngrafts were removed 15 and 30 min and 1, 4 and 24 h after reperfusion. MRI-imaging revealed significantly increased T2 relaxation times 1 and 4 h after reperfusion, indicating myocardial edema [18]. Permeability disturbance was normalized after 24 h of reperfusion (Figure 2A). Immunohistochemical analysis revealed that the peak influx of neutrophils occurred 4 h after reperfusion (Figure 2B, p < 0.05), while we observed no significant changes in the influx of macrophages (Figure 2C). Thus, our heart syngeneic transplantation model produced a transient IRI.

Figure 2.

HIF-1α expression sustained in SMC and CMC suffering from IRI. Edema in cardiac allografts was determined noninvasively by MRI by analyzing the myocardial T2 value 1, 4 and 24 h after reperfusion (A). The influx of neutrophils (B) and macrophages (C) during repefusion. The kinetics of HIF-1α protein during reperfusion, measured by ELISA (D). The number of CAIX+ PCV during reperfusion (E-G). HIF-1α immunoreactivity in PCV (H), arterial EC (I), SMC (J) and cardiomyocytes (K) in healthy (white) and injured (black) myocardial area of the syngraft. Representative photomicrographs of HIF-1α immunoreactivity in healthy and injured parenchyma (L-O, arrows) and arteries (P-S, arrows). Infiltrating PMN in syngraft (O, double arrows). *p > 0.01 **p > 0.05 versus normal nontransplanted DA rat heart, unless otherwise stated. n = 3/group. The red dashed line (A, D, E, H–K) equals levels in normal nontransplanted DA rat heart. Scale bars (F,G, L-S) are 20 μm.

ELISA revealed that nuclear HIF-1α protein was significantly increased 24 h after reperfusion, when compared to normal nontransplanted DA hearts (Figure 2D, p < 0.05). As a marker of transcriptional activity of HIF-1, we used protein expression levels of carbonic anhydrase IX (CAIX), which is directly downstream of HIF-1. Immunohistochemistry revealed that IRI time dependently increased the density of CAIX+ PCV (Figure 2E–G, p < 0.05).

To investigate whether HIF-1α expression is altered in the areas undergoing IRI in syngrafts, we determined HIF-1α immunoreactivity separately in the healthy myocardium, and in the areas with increased myocardial edema, hemorrhage and infiltration of inflammatory cells. Healthy and injured areas were compared to the values of normal heart. In the nuclei of EC in PCV, HIF-1α was not significantly upregulated during reperfusion (Figure 2H). In the nuclei of EC in arteries, we observed significantly increased HIF-1α immunoreactivity in both healthy and injured areas only 15 min after reperfusion, but not thereafter (Figure 2I, p < 0.01). In the nuclei of SMC in the coronary arteries, immunoreactivity of HIF-1α was significantly elevated in the areas undergoing IRI 15 min and 4 h after reperfusion (Figure 2J, p < 0.05). In the nuclei of cardiomyocytes, HIF-1α immunoreactivity was significantly elevated in IRI areas at 15 min, and at 1, 4 and 24 h (Figure 2K, p < 0.05). Representative photomicrographs of HIF-1α immunoreactivity in healthy myocardium and in the areas suffering from ischemia-reperfusion injury are shown in Figure 2L–S.

Ischemia reperfusion induces transient upregulation of endogenous protective genes downstream of HIF-1α in cardiac syngrafts

The mRNA levels of HIF-1α and some of its downstream genes were analyzed by quantitative real-time RT-PCR in cardiac syngrafts 0, 1, 4 and 24 h after reperfusion. No significant difference was observed in the mRNA expression of HIF-1α (Figure 3A). The mRNA levels of cardioprotective downstream genes of HIF-1, such as HO-1 (Figure 3B, p < 0.05) and Hsp27, however (Figure 3C, p < 0.05), were significantly increased during the reperfusion. In addition, we observed upregulation of transforming growth factor-β (Figure 3D, p < 0.05), and downregulation of proinflammatory transcription factor nuclear factor-κB (NF-κB) (Figure 3E, p < 0.05) and VEGF (Figure 3F, p <0.05).

Figure 3.

Cardioprotective HIF-1 downstream genes are expressed during myocardial reperfusion. mRNA levels of HIF-1 and various HIF-1 downstream genes during reperfusion (A–F). Red dashed line represents levels in normal nontransplanted DA heart. *p > 0.05 versus normal nontransplanted DA rat heart, unless otherwise stated. n = 3/group. The red dashed line (A–F) equals levels in normal nontransplanted DA rat heart.

Donor pretreatment with FG-4497 enhances IRI, while recipient pretreatment decreases ED1+ macrophages influx in cardiac allografts

Next, we investigated whether 4 h pretreatment of either donors or recipients with FG-4497 affects the degree of IRI in rat cardiac allografts 6 h after transplantation or the allograft long-term survival. The donors were treated 4 h before graft removal, and the recipients 4 h before heart transplantation. The effect of donor FG-4497 pretreatment was also investigated in syngrafts.

To confirm the efficacy of FG-4497, we initially treated rat SMC cultures with FG-4497 or vehicle. We observed intensive HIF-1α immunoreactivity in rat SMC cultures treated with FG-4497, while vehicle treatment did not induce any HIF-1α immunoreactivity in SMC (Figure 4A). mRNA analysis of FG-4497 treated SMC showed that downstream genes of HIF-1 such as HO-1 and ET-1 were upregulated (Figure 4B). The hearts from FG-4497 pretreated rats that were subjected to 4 h cold ischemia showed significantly upregulated HIF-1 downstream genes AngPTL4 (Figure 4C, p < 0.01) and HO-1 (Figure 4D, p < 0.01) mRNA levels. mRNA levels of angiopoietin-2 and iNOS were unaltered (Figure 4E and F).

Figure 4.

The efficacy of FG-4497 on activation of HIF-1 pathway. Nuclear HIF-1α immunoreactivity in rat SMC culture 24 h after adding 50 μM FG-4497 or vehicle (A). Relative mRNA levels of HIF-1 downstream genes HO-1 and ET-1 in SMC culture 72 h after adding 50 μM FG-4497 or vehicle (B). Relative mRNA levels of HIF-1 downstream genes in FG-4497 pretreated nontransplanted DA hearts or untreated hearts subjected ex vivo to 4-h cold preservation (C-F). White, control; black FG-4497 pretreatment. *p > 0.01 **p > 0.05. Scale bars (A) are 20 μm.

Analysis of cardiac allografts 6 h after transplantation indicated that recipient pretreatment significantly decreased intragraft density of ED1+ macrophages, while no significant differences were observed in the number of graft-infiltrating MPO+ neutrophils or CD4+ and CD8+ T cells. However, donor pretreatment enhanced serum levels of cardiac troponin T, an indicator of cardiomyocyte injury in cardiac allografts (Figure 5B, p < 0.05). Also, MCP-1, a chemokine guiding inflammatory cells to the site of injury (Figure 5C, p < 0.05), and the number of apoptotic cardiomyocytes, based on terminal deoxynucleotidyl transferase dUTP nick end labeling assay (TUNEL), were increased in the donor pretreatment group 6 h after transplantation (Figure 5D, p < 0.05). As EPO—a vascular hormone principally produced by kidneys and regulated by HIFs—is shown to have cardioprotective properties, we measured the serum EPO levels of the recipients, and found them to be significantly increased only in the recipient pretreatment group (Figure 5E, p < 0.05). FG-4497 pretreatment of the recipient mildly increased the long-term allograft survival (Figure 5F).

Figure 5.

The effect of donor and recipient FG-4497 pretreatment on ischemia-reperfusion and long-term allograft survival. The allograft infiltrating ED1+ macrophages, neutrophil, CD4+ T cells, and CD8+ T cells 6 h after reperfusion (A). Allograft serum cardiac troponin T (B), allograft serum MCP-1 (C), allograft serum EPO (D), the number of apoptotic cardiomyocyte nuclei in cardiac allografts based on TUNEL assay (E). Allograft survival (F). The syngraft infiltrating ED1+ macrophages, neutrophil, CD4+ T cells, and CD8+ T cells 6 h after reperfusion (G). Syngraft serum cardiac troponin T (H). *p > 0.01, **p > 0.05. White, control (n = 10); gray, donor pretreatment (n = 7); black, recipient pretreatment (n = 9); dense pattern, syngenic control (n = 5); thin pattern, syngenic donor pretreatment (n = 5).

In syngrafts, donor pretreatment did not alter the infiltration of inflammatory cells or serum cardiac troponin T levels. (Figure 5G and H).

In allografts, donor pretreatment enhanced the mRNA levels of innate immune ligand hyaluronan synthase 1, receptor TLR4, transcription factor NF-κB, proinflammatory cytokines IL-6 and IL-12p35, cytokine-induced neutrophil chemoattractant-1 (rat equivivalent of IL-8), vascular permeability factor VEGF, adhesion molecule ICAM and proapoptotic factor BAX in cardiac allografts 6 h after transplantation (Figure 6A–I, p <0 .05). Similar upregulation of innate immune genes was observed in the donor FG-4497 pretreatment group in syngrafts (Supporting Figure 1S A–I, p > 0.05).

Figure 6.

Innate immune genes are activated during reperfusion in rat cardiac allografts after FG-4497 donor pretreatment. Relative mRNA levels after FG-4497 pretreatment 6 h after heart transplantation in rat cardiac allografts (A-I). White, control; gray, donor FG-4497 pretreatment; black, recipient FG-4497 pretreatment *p> 0.05.

FG-4497 stimulation shifted macrophages into antiinflammatory phenotype, while human microvascular endothelial cells produced vasoconstrictive factors in vitro

We next investigated which potentially protective or unfavorable HIF-1 downstream genes are activated by FG-4497 treatment in vitro. mRNA extracted from FG-4497 stimulated macrophages showed that the levels of endothelium protective VEGF [19] and antiinflammatory, antiapoptotic and antiproliferative heme oxygenese-1 [20] were significantly upregulated, while no change was observed in the levels of proinflammatory TNF-α (Figure 7A). Furthermore, these macrophages also produced significant amounts of VEGF protein (Figure 7A). Interestingly, in human microvascular endothelial cells mRNA levels of vasoconstrictive ET-1 were increased after FG-4497 stimulation in normoxia and in hypoxia, while mRNA levels of iNOS were upregulated after FG-4497 stimulation only in normoxia, but not in hypoxia (Figure 7B).

Figure 7.

HIF-1 downstream gene expression favors antiinflammatory phenotype in macrophages (A) and vasoconstriction in human venous endothelial cell (B) after FG-4497 treatment. mRNA results are normalized to rat 18s ribosomal rna or human 18s Rrna, respectively. White, control; black FG-4497 treatment. *p > 0.05.


Here, we show that cold and warm ischemia of the heart transplant activated HIF-1α in a time- and temperature-dependent fashion. During reperfusion of the heart transplant, HIF-1α protein stability persisted in the injured areas, while it rapidly disappeared in the area without histological evidence of IRI. Unexpectedly, pharmacological preconditioning of the donor with the HIF-stabilizing prolyl-hydroxylase inhibitor FG-4497 did not protect against cardiac allograft IRI. Conversely, pharmacological preconditioning of the recipient decreased the infiltration of macrophages, and mildly improved the long-term allograft survival.

Generally HIF-1 is believed to act as an adaptive and survival factor for the tissue undergoing ischemic injury [21, 22], but some recent findings challenge this [23, 24]. Particularly in response to prolonged ischemic stress, HIF-1 may start to exert deleterious effects such as increased apoptosis [25, 26] and enhanced inflammatory processes [21, 27]. Inflammatory response may begin during normoxic activation of HIF before the tissue becomes hypoxic [28]. In hypoxia, HIF-1α levels accumulate and trigger the expression of genes involved in glycolysis, glucose metabolism, mitochondrial function, cell survival, apoptosis and resistance to oxidative stress [29]. Because HIF-1 regulates a variety of cellular functions, the balance between potentially beneficial and harmful genes eventually determines the final outcome.

We found that HIF-1α protein stability increased time and temperature dependently in all cardiac cell types during ischemia of the heart transplant. Interestingly, increased HIF-1α protein stability after short warm ischemia paralleled that of long cold ischemia, highlighting the importance of warm ischemia in HIF-1α activation. This is in accordance with a previous report showing that HIF-1α protein is effectively stabilized during warm acclimation [30]. Regarding myocardial injury in the rat, Bigaud et al. showed that creatine kinase release after 40 min of warm ischemia exceeds that of 6 h of cold ischemia in the experimental Langendorff model [31]. Together these findings imply that stabilization of HIF-1α may be related to the extent of myocardial injury. However, we found that HIF-1α protein expression persisted in the myocardium and in the SMC of coronary arteries in heart transplants suffering from IRI. This was linked to the mRNA upregulation of HO-1 and Hsp27 in heart transplants after reperfusion. The expression of these cardioprotective genes is regulated by HIF-1 [32, 33], and therefore we hypothesized that HIF-1 could be involved in the development and the resolution of myocardial IRI.

The interest in pharmacological agents mimicking IPC or rIPC is vast, and the effects of CoCl2 [34], iron chelators [35], volatile anesthetics [36] and prolyl hydroxylase inhibitors [37] have proved beneficial in several studies on myocardial ischemia and reperfusion. In these studies, the activation of HIF-1 pathway has been indispensable. However, the ischemia times have been relatively short (usually less than 1 h), and none of these studies examined myocardial ischemic preconditioning in allogenic setting. Contrary to our initial hypothesis and to previous results in rat kidney allografts [38], our results showed that under the conditions of this study, pharmacological HIF preconditioning of the donor enhanced acute myocardial injury, cardiomyocyte apoptosis and activation of innate immune responses in cardiac allografts 6 h after reperfusion. In contrast, Bernardt et al. recently described that donor treatment with a single dose of FG-4497 i.v. 6 h before the onset of 24-h cold ischemia ameliorated acute kidney injury, early mortality and improved the long-term survival of rat kidney allografts [38]. Our contradictory findings need to be addressed. The route of administration and time scale differ and that may affect the preconditioning potential obtained in the transplant donor. The normal physiology of the kidney and the heart is very different. Myocardium is constantly well oxygenated. However, vigorous oxygen gradient exists between the medulla and the cortex of kidney and the renal medulla has the lowest physiologic PaO2 compared with any other organ despite the fact that the kidney receives the largest fraction of the cardiac output. The kidney is generally more resistant to ischemia than the heart, and the mean duration of the total ischemia in the transplantation clinics is 2.8 h for hearts [39] and 22.6 h for kidneys [40]. Finally, we may speculate that as FG-4497 is effectively stabilizing both HIF isoforms (HIF-1 and HIF-2), the donor FG-4497 pretreatment could have different effects in different organs due to induction of tissue-specific HIF-1 or HIF-2 target genes. The distribution pattern of the two HIF-α isoforms differs between heart and kidney as in the heart the expression of the both transcription factors overlaps to a large extent, while in the kidney HIF-1 is expressed in tubular and HIF-2 in interstitial and endothelial cells [41, 42]. Specifically, HIF-mediated EPO may provide autocrine renoprotection in kidney allografts whereas similar protection may not be seen in cardiac allografts as kidney is the primary source of EPO. EPO may also facilitate cardioprotection via remote renal preconditioning [43], but this did not happen in our study. Enhanced purinergic signaling has been shown to involve in HIF-mediated ischemic preconditioning in the heart. We could not show this under the conditions of our study (data not shown) [44].

Scientific papers suggesting a detrimental role for HIF-1α in myocardial injury have also emerged. In support of our findings, activation of HIF-1α during IRI in the rat heart leads to iron and transferrin accumulation, which exacerbates oxidative damage [45]. Sustained HIF-1 activation causes enhanced premature mortality due to venous congestion and dilated cardiomyopathy [46], and prolonged HIF-1α activitity in the heart led to maladaptation, cardiac degeneration and progression of heart failure [47].

In transplantation studies, donor myocardial HIF-1α is an independent predictor of cardiac allograft dysfunction after heart transplantation [48] and associates with renal allograft rejection 3 months after the transplantation [49]. Furthermore, it was recently shown that prolonged intestinal HIF-1 activation is a proximal regulator of IRI-induced gut mucosal injury and gut-induced lung injury [23]. Kanna et al. also reported that similarly to cardiac allografts, IRI of the gut was directly proportional to the duration of the ischemia [24, 50, 51]. Rosenberger et al. previously speculated that there appears to be a ‘window of opportunity’ for a HIF-dependent transcriptional response, and if the severity of hypoxia exceeds this range, the adaptive response fails and injury may become irreversible [41].

We hypothesize that prolonged ischemia with prior pharmacological HIF-1 activation in the donor might cause detrimental microvascular dysfunction, which could lead to increased acute myocardial injury during reperfusion. The severe imbalance in metabolic demand and supply, as in transplantation setting, may cause microvascular dysfunction [52]. qPCR analysis of the allograft in donor pretreatment rats show vast upregulation of the innate immunity cascade from the ligands of the TLRs to the proinflammatory cytokines. In kidney transplant, the expression of TLR4 correlated with the degree of ischemic injury [53]. Similar to the findings with allografts, donor FG-4497 pretreatment increased the expression of innate immunity genes in syngrafts, while it had no effect on cardiomyocyte damage. These findings may thus highlight the importance of alloimmunity as an integral component of IRI in transplanted organs. El-Sawy et al. have previously shown that innate immune response in syngrafts resolves and that CD8+ T cells attack the allogenic vascular endothelium within hours of reperfusion [54]. In addition, the presence of alloimmunity in our transplantation model may at least in part explain why previous nontransplant experiments have shown that pharmacological HIF-1 preconditioning results in myocardial protection.

Furthermore, we treated human microvascular endothelial cells with FG-4497 and exposed them to normoxic and hypoxic conditions. In human microvascular endothelial cell culture, we found that pharmacological HIF-1 activation upregulated endothelial cell expression of ET-1 mRNA in normoxia, while in hypoxia ET-1 mRNA was highly upregulated irrespective of pharmacological HIF-1 stabilization. On the contrary, FG-4497 increased normoxic iNOS mRNA, while in hypoxia there was no effect. Endothelium is an important source of vasoactive substances that affect nearby vascular smooth muscle cells in a paracrine fashion. Under the conditions of this study, the balance of genes examined affecting the vascular tone favors vasoconstriction.

To conclude, our results show that ischemia and IRI activate the HIF-1 pathway in the heart transplant microenvironment leading to transient induction of vasoactive genes downstream of HIF-1. Thus, HIF-1 activation is directly proportional to the myocardial injury, not as a causing agent but as a consequence. Unexpectedly, pharmacological HIF-α preconditioning of the donor for 4 h enhanced myocardial ischemia reperfusion injury and innate immune activation in cardiac allografts. In contrast, preconditioning of the recipient for 4 h reduced cardiac allograft inflammation, and mildly improved the long-term allograft survival. These results suggest that HIF stabilization in organ transplant donors would not provide clinical benefit with respect to cardiac allografts, although previous studies suggest that the opposite may be true for renal allografts.


We thank Eeva Rouvinen, RN. Lab., Jaana Komulainen, RN. Lab. Ralica Arnaudova, Msc and Satu Lehti, MSc, for their excellent technical assistance and Usama Abo-Ramadan, PhD, from Experimental MRI Laboratory, Department of Neurology, Helsinki University Central Hospital, Helsinki, Finland, for MRI-imaging.

This study was financed by a grant from The Roche Organ Transplantation Research Foundation, Basel, Switzerland. Additional grant support from the Academy of Finland, Helsinki University Central Hospital Research Funds, the Sigrid Juselius Foundation, Finnish Cultural Foundation, Finnish Foundation of Cardiovascular Research, Päivikki ja Sakari Sohlberg Foundation, the Aarne Koskelo Foundation, Research Foundation of the University of Helsinki, the Aarne and Aili Turunen Foundation, Finska Läkaresällskapet, the Research and Science Foundation of Farmos, The Finnish Medical Foundation, Biomedicum Foundation, Paavo Nurmi Foundation, all from Helsinki, Finland, Emil Aaltonen Foundation from Tampere, Finland, and the Karoliina and Kustaa Lindeman Foundation from Jämsä, Finland, is also acknowledged.


The authors of this manuscript have conflicts of interest to disclose as described by the American Journal of Transplantation. G.W., L.A.F., and M.A. are employees of and own equity in FibroGen Inc., which owns the commercial rights to FG-4497.